• Keine Ergebnisse gefunden

Structure and dynamics of the aggregation mechanism of the Parkinson´s disease-associated protein alpha-synuclein

N/A
N/A
Protected

Academic year: 2022

Aktie "Structure and dynamics of the aggregation mechanism of the Parkinson´s disease-associated protein alpha-synuclein"

Copied!
235
0
0

Wird geladen.... (Jetzt Volltext ansehen)

Volltext

(1)

Structure and dynamics of the aggregation mechanism of the Parkinson´s disease-associated protein

α-synuclein

Dissertation

zur Erlangung des Doktorgrades

der Mathematisch-Naturwissenschaftlichen Fakultäten der Georg-August-Universität zu Göttingen

vorgelegt von Carlos Walter Bertoncini aus Santa Fe, Argentinien

Göttingen, 2006

(2)

D7

Referent: Prof. Ralf Ficner Korreferent: Prof. Reinhard Jahn Tag der mündlichen Prüfung: 05.07.06

(3)

In Parkinson’s disease (PD), intracellular neuronal inclusions containing amyloid-like aggregates of the protein α-synuclein (αS) are deposited in the pigmented nuclei of the brainstem. The mechanisms underlying the structural transition of innocuous, presumably natively unfolded αS to oligomeric neurotoxic forms of the protein are largely unknown.

The major aim of this thesis has been the characterization of the ensemble of conformers that αS populates in its monomeric native state and the early transitions that lead to protein oligomerization, by the use of nuclear magnetic resonance (NMR) spectroscopy.

The conformational flexibility inherent to this kind of proteins places them beyond the reach of classical structural biology, and a special set of NMR-based experiments had to be implemented in order to study the native soluble state of αS, namely paramagnetic relaxation enhancement (PRE) from nitroxide spin labels and residual dipolar couplings (RDCs).

Our results show evidence that monomeric αS assumes conformations that are stabilized by long-range interactions and act to inhibit aggregation. As probed by RDCs, these auto-inhibitory conformations are formed on a ns to μs timescale that is precisely that in which secondary structure elements form during folding. In addition, PRE-derived distance restraints have been employed to derive a low resolution model for the ensemble of structures of αS compatible with the experimental findings.

The conformations populated by αS under disease-state circumstances were further investigated. Missense mutations linked to early onset PD, and environmental conditions that promote αS aggregation were found to release the inherent tertiary structure of the protein.

Thus mutant or ligand bound αS overcomes more easily the energetic barrier for self- association, leading to an increased tendency to oligomerize.

The homologous protein βS, which is proposed to inhibit the toxicity of αS, has been also characterized by means of high resolution NMR. It was found that the conformations populated by this protein do not account for long range interactions as αS does, but a higher degree of residual structure is attained, likely polyproline II extended conformations.

Finally, the binding of divalent transition metal cations to αS was studied by a set of spectroscopic techniques. It was found that among several transition metals, Cu(II) is strongly bound by the N-terminus of the protein with an affinity of ~ 100 nM, and the complex is more prone to aggregate that the free protein. Other metals as Fe(II), Mn (II), Ni(II) and Co(II) do not influence protein aggregation since they bind to the C-terminus of the protein

(4)

with 104 lower affinities. Our results suggest that impairment of Cu(II) homeostasis links the three major amyloid neurodegenerative disorders Alzheimer’s, Prion and Parkinson’s disease.

From the therapeutic point of view, it is foreseen that the reinforcement of these native, auto-inhibitory, long-range interactions in αS may be a (the) key target of new pharmacological agents designed to impede or even reverse aggregate formation in Parkinson’s disease. Conformationally altered αS may also constitute a general molecular mechanism underlying the induction of PD by both environmental and genetic conditions.

Thus, agents specifically designed to stabilize the native state of αS may also prove useful in impeding or reversing its pathologic aggregation in familial forms of PD.

(5)

1. Introduction... 3

1.1. Protein folding... 3

1.1.1. The energy landscape of protein folding... 3

1.1.2. Protein folding in the cell... 5

1.1.3. Quality control mechanisms of protein folding... 7

1.2. Protein misfolding and disease... 8

1.2.1. Amyloid diseases... 10

1.2.2. Molecular basis of amyloid formation... 10

1.2.3. Conformational plasticity in amyloid formation... 13

1.3. Protein misfolding in Parkinson’s disease... 16

1.3.1. α-Synuclein aggregation is linked to Parkinson’s disease... 17

1.3.2. The physiological role of α-Synuclein... 19

1.3.3. The native unfolded state of α-synuclein... 19

1.3.4. Ligand induced aggregation of α-synuclein... 21

1.3.5. Polycation-induced α-synuclein fibrillation... 21

1.3.6. Metal induced α-synuclein fibrillation... 24

1.4. The unfolded state of proteins... 25

1.4.1. Intrinsic conformational restrictions in the unfolded state... 25

1.4.2. Structural studies on the unfolded state of proteins... 27

1.4.3. Structural studies on α-synuclein... 29

2. Aims of the thesis... 33

3. Materials and methods.... 37

3.1 Materials.... 37

3.1.1. Equipment.... 37

3.1.2. Bacterial strains.... 37

3.1.3. Reagents.... 38

3.2. Methods... 38

3.2.1. Molecular biology.... 38

3.2.1.1. αS-containing plasmids.... 38

3.2.1.2. βS-containing plasmids.... 39

3.2.1.3. Construction of αS and βS Cys-containing mutants... 39

3.2.2. Protein biochemistry... 41

(6)

3.2.2.1. Expression and purification of αS and βS... 41

3.2.2.2. Synthesis of C-terminal peptides of αS.... 43

3.2.2.3. Spin-labelling of proteins.... 43

3.2.2.4. Alignment of αS in anisotropic media.... 44

3.2.2.5. Protein aggregation assays... 45

3.2.2.6. Metal content determination in αS-metal(II) complexes.... 46

3.2.2.7. Equilibrium dialysis assays... 46

3.2.2.8. Chemical modification of His residue in αS... 47

3.2.3. Spectroscopic determinations.... 47

3.2.3.1. Absorption and CD spectroscopy.... 47

3.2.3.2. EPR spectroscopy... 47

3.2.3.3. NMR spectroscopy.... 47

3.3. Miscellaneous... 60

3.3.1. Calculation of distance restraints from PRE.... 60

3.3.2. Structure determination and analysis.... 60

3.3.3. Electron microscopy... 61

4. Results. Chapter I: “NMR identifies long-range auto-inhibitory interactions in the native state of αS”.... 65

4.1 Heteronuclear 2D NMR spectroscopy probes conformational transitions on αS.65 4.2. Residual dipolar couplings evidence residual structure in native αS.... 72

4.3 Paramagnetic resonance enhancement on αS detects long range interactions..... 76

4.4. A conformational ensemble representative of the native state of αS... 80

4.5. Polyamine binding releases long-range interactions in αS.... 83

4.6. The residual structure of native αS is lost at elevated temperatures... 86

4.7. αS populates a native-like destabilized conformation at pH 6.5... 89

5. Results. Chapter II: “PD-linked familial mutants of αS have a destabilized conformation”.... 93

5.1. Conformations of familial mutants of αS studied by heteronuclear 2D NMR... 93

5.2 The A30P and A53T familial mutations perturb tertiary interactions in the native state of αS.... 96

5.3. A30P and A53T mutations destabilize αS.... 100

5.4. Synergistic long range interactions occur in the native state of αS.... 104

5.5. Competition of electrostatic and hydrophobic interactions in the native state of αS... 106

6. Results. Chapter III: “Structural characterization of the PD-associated protein β- Synuclein.”... 111

(7)

analysis... 116

6.4. Residual structure in βS probed by dipolar couplings.... 122

6.5. Residual structure in βS is locally encoded... 124

6.6. Polyproline II structure is evidenced at the C-terminus of βS... 126

7. Results. Chapter IV: “Structural basis of metal binding to α-synuclein”.... 131

7.1. Cu(II) binding to αS promotes protein aggregation... 131

7.2. Quantitative assessment of Cu(II) binding to αS.... 133

7.3. Dissecting domain contributors to Cu (II) binding to αS... 136

7.4. Mapping Cu(II) binding interfaces in αS by heteronuclear NMR.... 137

7.5. Cu(II) coordination in the complex with α-synuclein studied by EPR... 141

7.6. Paramagnetic metal ions bind to αS with different affinities.... 143

7.7. NMR characterization of the interaction between divalent metal ions and αS.. 145

8. Discussion... 153

8.1. Auto-inhibitory long range interactions in the native state of αS... 153

8.2. Residual structure in the ensemble of conformers populated by αS.... 156

8.3. Dissecting the nature of long range interactions probed by RDCs in αS... 158

8.3. RDCs probe slow conformational dynamics in αS... 161

8.4. Stabilization of auto inhibitory interactions in αS may inhibit amyloid protein deposition... 164

8.5. Redistribution of the ensemble of αS conformers may underlie toxic gain-of- function in A30P and A53T genetic mutants... 165

8.6. Characterization of the structural basis of metal-αS complexes.... 166

8.7. Cu(II) binding is a biological link among amyloid diseases.... 172

9. Conclusions... 177

10. References... 181

11. Appendix... 203

11.1 Fluorescence spectroscopy and microscopy to study α-synuclein fibrillation... 203

11.2 Backbone assignment of the native soluble state of the protein βS... 213

Acknowledgments... 219

Publications... 221

Lebenslauf... 225

(8)
(9)

Abbreviations Αβ Αmyloid-β peptide

AD Alzheimer’s disease

APP Amyloid precursor protein

αS α-synuclein

ATP Adenosine -5’-Triphosphate

B0 External magnetic field βS β-synuclein

CD Circular dichroism

DEPC Diethylpyrocarbonate

DLB Dementia with Lewy bodies

CMA Chaperone-mediated autophagy

DNA Desoxyribonucleic acid

1DNH N-H residual dipolar coupling

DSSE Doublet Separated Sensitivity Enhanced DTNB 5,5’-Dithiobis(2-nitrobenzoic acid)

DTT Dithiothreitol

EDTA Ethylenediamine tetraacetic acid EPR Electron paramagnetic resonance

ER Endoplasmic reticulum

FRET Förster resonance energy transfer

FTIR Fourier-transform infrared spectroscopy γS γ-synuclein

HMQC Heteronuclear double quantum coherence HSQC Heteronuclear single quantum coherence HSPs Heat shock proteins

INEPT Insensitive nuclei enhanced by polarization transfer

IPAP In-Phase-Anti-Phase

IPTG Isopropyl-β-D-thiogalactopyranoside

1JNH N-H scalar coupling

3JHNHα HN-Hα scalar coupling

LBs Lewy bodies

(10)

MPP+ 1-Methyl-4-phenylpyridinium ms Milliseconds

μs Microseconds

MTSL (1-oxy-2,2,5,5-tetramethyl-D-pyrroline-3-methyl)- methanethiosulfonate

NAC Non-Aβ Component of Alzheimer’s disease amyloid plaques NMR Nuclear magnetic resonance

NOE Nuclear overhausser effect

ns Nanoseconds

PAR 4-(2-pyridylazo)resorcinol

PCR Polymerase chain reaction

PD Parkinson’s disease

PFG-NMR Pulse field gradient-NMR

PG-SLED Pulse gradient stimulated echo longitudinal encode-decode PMSF Phenylmethylsulfonyl fluoride

PII Polyproline II

PRE Paramagnetic relaxation enhancement

PrP Prion protein

ps Picoseconds

RDCs Residual dipolar coupling Rg Radius of gyration

Rh Hydrodynamic radius

RT-PCR Reverse transcription-polymerase chain reaction

SAXS X-ray scattering

SDS Sodium Dodecyl Sulphate

SDS-PAGE SDS-polyacrylamide gel electrophoresis

TF Trigger factor

Thio-T Thioflavin-T

Ub Ubiquitin

UCH-L1 Ubiquitin Carboxy-Terminal Hydrolase-L1

UPS Ubiquitin-proteasomal system

wt Wild type

(11)

I I n n t t r r o o d d u u c c t t i i o o n n

(12)
(13)

1. Introduction 1.1. Protein folding

Organisms have evolved such as the native states of proteins are required to perform diverse biochemical functions, ranging from mere pillars of the overall cellular architecture, like actins and tubulin, up to exquisitely complex molecular machines, as the DNA and RNA polymerases. In all cases adopting a proper structure is essential if the proteins are to carry out their biological function. Failing to do so has often critical consequences for cellular homeostasis (Dobson, 2003).

The mechanism by which a polypeptide chain attains its unique native three dimensional structure is known as protein folding. Inside a cell proteins fold during, or immediately after, their synthesis in response to the crowding nature of the media and the concerted action of chaperones (Young et al., 2004). Nevertheless the fold of a protein appears to be solely encoded in the primary amino acid sequence, inasmuch as polypeptide chains fold in vitro in the absence of any auxiliary factors (Anfinsen, 1973; Dobson and Karplus, 1999). Indeed, the self assembly capability of proteins has allowed the occurrence of the early biologically-driven chemical processes, and the stability of such catalytically- competent polypeptide-based systems is believed to have been crucial for evolution to have taken place.

1.1.1. The energy landscape of protein folding

An unbiased search for the most stable structure of a protein would demand the population of all possible conformations that the polypeptide chain could adopt, until the state with the lowest energy is attained. As the average size of naturally occurring proteins is about 300 amino acids, it would take an astronomical amount of time for a polypeptide to fold via a random search of the conformations (Fersht, 1999). However, most proteins fold in the order of μs to ms, suggesting a preferential population of conformational states, as already proposed by Levinthal in the late ´60s (Dill and Chan, 1997).

Several models have emerged to explain the existence of folding pathways, from the framework model according to which secondary structure forms locally followed by collision of the folded segments, to the nucleation model in which folding is initiated locally and propagated through the chain. A third model, the hydrophobic collapse model, hypothesizes that the polypeptide would collapse rapidly around its hydrophobic side chains and then rearrange from the restricted conformational space that is then sampled (Fersht, 1999). All

(14)

three models are reasonable and fit experimental data, such that a single generic mechanism for protein folding may not exist in nature.

A conceptual mechanism for understanding how a protein folds while avoiding the Levinthal paradox is a folding funnel, described by the most probable pathways leading to the adoption of the unique native structure of a protein (Wolynes et al., 1995; Dill and Chan, 1997; Dinner et al., 2000). In a folding funnel, as schematized on figure 1.1, the unfolded state possesses the highest free energy and comprises an ensemble of multiple conformations constituting the starting point of the folding pathway. The fast dynamics and low conformational restriction that characterize this state allow the occurrence of interactions between different parts of the chain, and, among them, native-like contacts involving key residues are energetically favored. This causes a fast collapse of the polypeptide chain, often driven by the formation of hydrophobic clusters decreasing the free energy of the system.

Thus proteins may fold from the random state by collapsing and reconfiguring (Fersht, 1999).

A loosely collapsed state with fluctuating tertiary interactions and very weak secondary structure, known as the “molten globule”, may be observed, and some times isolated.

Figure 1.1. Landscape for protein folding. Theoretical folding surface for a polypeptide chain. One folding pathway is indicated by the continuous arrowed line, where the rapid collapse of the chain is represented by the population of the first valley, which makes that only few particular conformations overtake the higher free energy of the transition state. After the saddle point is reached folding readily occurs downhill.

Adapted from Dinner et al. (Dinner et al., 2000).

(15)

Neighboring residues start then to populate conformations with increasing amount of stable secondary structure, with turns and helices forming more rapidly than beta sheets (Eaton et al., 1998; Plaxco et al., 1998; Mayor et al., 2003). The population of a defined secondary structure requires that many residues concomitantly adopt concerted backbone conformations capable of forming hydrogen bonds between amide and carbonyl groups of the main chain, which further reduce the degree of freedom of the system. In many cases, depending on the mechanism by which the protein folds, a relatively stable transition state is populated, which adopts an overall three-dimensional shape that closely resembles the one of the native state of the protein (Vendruscolo et al., 2003). After the transition state the protein folds downhill until it attains the single unique state with the lowest energy.

In summary, the native state of a protein is achieved through a precise interplay of both short- and long-range interactions, determining that only a small number of conformations need to be sampled, thereby speeding up the process.

1.1.2. Protein folding in the cell

In a cellular environment protein folding occurs co-translationally in the cytoplasm, mainly due to the slow synthesis rate by the ribosome, of ~ 8 amino acid residues per second. The concerted effect of chaperones prevents hydrophobic stretches from newly synthesized polypeptides to be exposed and to self-interact, avoiding protein aggregation (Hartl and Hayer-Hartl, 2002). Major cytosolic chaperones belong to the class of small heat shock proteins (HSPs), like the HSP60s (GroEL-GroES, CCT/TRiC), HSP70s, HSP90s and HSP100s, and another important class is represented by foldases, like the peptidyl-prolyl cis- trans isomerases, which isomerases peptidyl-prolyl bonds in proteins (Bukau and Horwich, 1998).

Insights into the mechanism of protein folding in vivo were recently provided by high resolution structural studies of bacterial chaperones, but mammalian systems are homologous and very likely to proceed similarly (Bukau and Horwich, 1998). In bacteria protein synthesis occurs at a speed of 20 residues per second, and the nascent polypeptide leaves the ribosome through a 10 nm exit tunnel placed in the large subunit. This hollow cage may accommodate up to 35 residues, but its narrow diameter (about 15 Å) impedes folding. To the end of this tunnel is recruited the ribosome-associated chaperone trigger factor (TF), a 48 kDa protein which provides a cage where hydrophobic residues of the nascent polypeptide are able to fold. TF possess high affinity for emerging exposed hydrophobic stretches but the interaction weakens as the protein folds and such residues are buried, leading to detachment from the

(16)

complex (Maier et al., 2005). Two other systems provide folding assistance either co- or post- translationally, the DnaK-DnaJ chaperone complex (HSP70 family) and the GroEL-GroES chaperonin (HSP60 family). In contrast to TF, these two mechanisms require ATP and co- chaperone systems. DnaK preferentially associates with elongating polypeptides larger than 20 to 30 kDa and thus acts on nascent chains subsequent to TF. It does not posses a cage, and rather facilitates the posttranslational folding of multidomain proteins through cycles of binding and release (Teter et al., 1999). The GroEL-GroES system forms a hydrophobic cage (also termed “Anfinsen cage”) in which non-native proteins up to 60 kDa can be encapsulated and are free to fold. Folding is allowed to proceed for 10 s, as timed by the hydrolysis of ATP (Hartl and Hayer-Hartl, 2002).

The concerted action of the three chaperone-based system ensures proper folding in the cell; however it remains to be determined which fraction of proteins does entirely rely on chaperones to fold. For example a recent quantitative assessment of the GroES-GroEL chaperone function in vivo suggests that it may only be involved in the folding of ~ 100 proteins in E. coli. (Kerner et al., 2005).

Proteins that are directed to the secretory pathway or to mithocondria do not fold in the cytoplasm since they have to traverse the phospholypidic bilayers, remaining thus unfolded until they reach their final destination. Protein folding in the endoplasmic reticulum (ER) is also assisted by chaperones of the HSP70 and HSP90 class, GrpE-like and DnaJ-like, among others, and foldases, lectines and N-linked oligosaccharide-modifying enzymes are also present to ensure that polypeptides adopt their proper structure. Importantly, prior to vesicular release, folded proteins suffer a tight quality check based on glycosylations and de- glycosilations by the proteins UGGT and BiP (Schroder and Kaufman, 2005).

In vitro, re-folding after denaturation seems to be successful only for small to medium-size proteins, and very often large multidomains proteins fail to re-fold, mainly due to intra- and inter-molecular interactions (Kiefhaber et al., 1991). The picture in vivo is very similar, where off-pathway conformations are more likely to be populated if chaperones would not protect hydrophobic patches and force nascent polypeptide domains to fold independently and in a subsequent manner. Thus, in kinetic terms, chaperones do not themselves increase the rate of folding, rather they foster the overall efficiency of the process by reducing the probability of competing reactions, in particular aggregation (Schmid et al., 1994). In addition, confinement in the chaperone cage may smooth the energy landscape of folding for some larger proteins, either by preventing the formation of certain kinetically

(17)

trapped intermediates or by facilitating their progression toward the compact, native state (Hartl and Hayer-Hartl, 2002). As a conclusion, the evidence points that in vivo the folding of a protein is also exclusively dictated by the amino acid code of the polypeptide chain, and thus the generic applicability of the principles of protein folding (Dobson, 2003).

1.1.3. Quality control mechanisms of protein folding

Cells possess quality control mechanisms that ensure a proper folding of proteins and facilitate removal of misfolded species, mainly based on the concerted action of chaperones and the ubiquitin-proteasomal system (UPS). Although the buffering capacity of chaperones may be sufficient for counterbalancing small changes in the amount of non-native proteins, it is often required to clear species that do not properly re-fold and may aggregate (Buchner, 1996). To achieve such a task, cells couple the chaperone system to proteasome-based pathways of protein degradation, like the chaperone BiP (HSP70 class), which binds to a mutant Prion protein and mediates its degradation by the proteasome (Jin et al., 2000). The proteasome is a multisubunit complex that manages protein turnover, either by direct interaction with the protein substrate or by recognition of ubiquitinated substrates. Three different proteasomes are found in mammalian cells, two which are ubiquitin-independent, the 20S and 26S, and one ubiquitin-dependent, the 26S, the latter being the main degradation pathway of the cell (Ciechanover, 2005).

Chaperones are not only able to protect proteins as they fold, but they also rescue misfolded and aggregated proteins, providing them a new chance to get properly folded (Hartl and Hayer-Hartl, 2002). In particular, the chaperone HSP100 is thought to participate in ATP-dependant dissagregation of proteins, and polyglutamine-expanded ataxin 1 misfolding and aggregation is suppressed by chaperones of the HSP70 family (Cummings et al., 1998).

Extracellular protein-rich depositions are also in part derived from the misfolding of polypeptides, and in the light of few molecular chaperones present outside the cell, extracellular membrane or secreted proteins undergo a tight quality check in the ER. The concerted action of glycosilases and deglycosilases determines the fate of proteins that cannot fold correctly (Schroder and Kaufman, 2005). Indeed, it was observed that a large fraction of newly synthesized proteins fail to pass the stringent checks at the ER, and are thus degraded by the proteasome (Schubert et al., 2000).

(18)

1.2. Protein misfolding and disease

The quality control mechanisms of protein folding discussed above are often overpowered in a disease state of the cell, increasing the chances for a polypeptide chain to adopt conformations that are not conductive to the native structure of proteins. Furthermore, feedback to such impairment of cell homeostasis will come from proteins that do not fold correctly, or do not remain correctly folded, since they are often prone to aggregate and to form toxic oligomers, which sustain proteasomal impairment (Dobson, 2003) .

For example, a primary trigger for proteasomal failure is a massive overloading of unfolded substrates, but UPS malfunction may also arise from deficiencies in ubiquitinating enzymes responsible for determining the fate of misfolded proteins (Ciechanover and Brundin, 2003). Ubiquitin-independent degradation via the proteasome is also heavily challenged upon stress situations or when chaperones are incapable of refolding proteins, and accumulated misfolded proteins, which are not able to be cleared under these conditions, are progressively deposited in protein-rich aggregates called aggregosomes (Johnston et al., 1998). Moreover, recent studies suggest that such protein aggregates cause cytotoxicity by inhibiting proteasomal and/or UPS function, which may be causative of neurodegenerative disorders such as Alzheimer’s, Parkinson´s, Huntington’s and Prion diseases (Bence et al., 2001; Ciechanover and Brundin, 2003). Indeed, impairment of proteasomal function in the substantia nigra has been described for Alzheimer’s and Parkinson´s patients (McNaught and Jenner, 2001; Keck et al., 2003). The mechanism by which this inhibitory effect takes place is largely unknown, but supports the idea of a generic toxic property of protein aggregates (Dobson, 2003).

In addition, the function of the ER is perturbed when the influx of unfolded polypeptides exceeds its folding capacity, which is often the case upon over-expression of proteins such as antithrombin III or blood coagulation factor VIII, which sequester chaperones and form protein aggregates. The ER normally responds by a complex signaling cascade known as the unfolded protein response, increasing the expression of chaperones and generally repressing transcription and translation (Schroder and Kaufman, 2005). Misfolded proteins in the ER are retro-translocated to the cytoplasm and degraded by the 26S proteasome (Mayer et al., 1998). Nevertheless, long-term abnormal accumulation of folding- incompetent proteins may produce oligomers resistant to proteasomal degradation, which completely disrupts the ER and activates apoptotic signaling pathways, leading to a disease- state of the cell (Davis et al., 1999; Nakagawa et al., 2000).

(19)

It is therefore not surprising that aberrations in the folding and abnormal deposition of certain proteins are now associated with a wide range of human diseases, collectively known as misfolding diseases (table 1.1). Such disorders may arise from the presence of mutations that impede a proper folding of the polypeptide chain or from stress factors that specifically trigger the misfolding of the protein, escaping control mechanisms and forming toxic species to cells (Dobson, 2004).

Table 1.1. Representative protein folding diseases. Selected disorders that are linked to the misfolding and aberrant deposition of proteins are listed. Proteins responsible of such deposits and the respective cellular compartment of folding are also depicted. Proteins displaying amyloid deposition in disease are highlighted in bold.

Adapted from Dobson, 2004 (Dobson, 2004).

A common denominator of many if not all of these diseases is the modern-man intervention. For example, spanning our life-term permits accumulating damage to cells reducing the stringency of quality control mechanisms, in particular for protein folding, facilitating the occurrence of Alzheimer’s (AD) and Parkinson´s disease (PD). Exposure to environmental toxins, like MTTP or Rotenone is another major cause of PD, and medical treatments as long-term dialysis unexpectedly cause deposition of proteins in liver or

(20)

osteoarticular tissue. The outburst of mad-cow disease during the last decade have also been a consequence of man’s intervention in the feeding of animals, allowing misfolded proteins to cross natural barriers. The current idea behind these afflictions is that the actions of modern man have radically altered the evolutionary pressure on proteins. Thus the challenge to science strategists is to device the means for circumventing interventions with deleterious effects while promoting those offering ecologically responsible benefits to the human race.

1.2.1. Amyloid diseases

A sub-set of these pathologies, including AD, PD, Prion diseases and late-onset diabetes, among others, are associated with the deposition of structurally defined protein aggregates in the tissue known as amyloid fibrils (Koo et al., 1999; Serpell, 2000). Amyloid in disease is generally defined to be extracellular, although intracellular structures sharing the same core structures are described in PD (Serpell et al., 2000).

Amyloid is defined in terms of empirical observations from X-ray fiber diffraction, electron microscopy, FTIR and specific chemical staining with dyes such as Congo Red and Thioflavine T. Thus to be classified as an amyloid protein, its deposits should be straight, unbranched, of about 10 nm in diameter, reach a μm in length, present a cross-β diffraction with two sharp reflections at 4.7 Å and 10 Å, and display green birefringence after staining with Congo Red.

The pathological hallmark of these conformational diseases is not constrained to the formation of amyloid-like fibrils. In several disorders, protein deposits are composed of amorphous aggregates, without micro- or macroscopical order. Similarly, stable soluble oligomers derived from the self-association of soluble misfolded species, could be the final product of the aggregation process, and the culprit of many of these disorders (Caughey and Lansbury, 2003).

1.2.2. Molecular basis of amyloid formation

In order to deposit in the form of ordered filamentous protein-rich aggregates, a dramatic change in the structure of a protein has to occur. A conformational change triggered on the polypeptide chain cause a transition from its natural soluble conformation towards a more insoluble state. Although more than 20 proteins are known to be involved with such pathological depositions, they do not share any sequence homology, and may be either rich in β-sheets, α-helix, or lack significant secondary structure in their native states (Uversky and Fink, 2004). However, independently of the originating protein, they all form a common

(21)

cross-β structure in which continuous β-sheets are formed with β-strands running perpendicular to the fibril axis (Makin and Serpell, 2005).

To reconcile this observation, it has been proposed that fibrillation may occur when the rigid native structure of a protein is destabilized, favoring a partial unfolding. Such partially or completely unstructured conformations of the polypeptide chains are particularly high aggregation-prone, presumably via specific non-native intermolecular interactions conductive for oligomerization (Uversky and Fink, 2004). Indeed, many of the pathogenic mutations that are associated with familial deposition diseases increase the population of partially unfolded states by decreasing the stability of the native state or reducing its global cooperativity, as observed for lysozyme (Booth et al., 1997; Dumoulin et al., 2005). For Transthyretine it has also been postulated that pathogenic mutants destabilize the native tetramer of the protein, favoring a kinetic partition towards the monomeric state which easily destabilizes and aggregates (Hammarstrom et al., 2002).

Conformations that lead to misfolded states of a protein are not dictated by the polypeptide sequence, and thus are not generally part of the folding funnel, which is sequence-specific, giving rise to a second energetic landscape known as misfolding pathway (Figure 2.1). Although misfolded structures possess a low free energy, and could even be more stable than the native state of the protein, the energetic barrier dictates that the unfolded state has to be attained earlier than such conformations are populated (Uversky and Fink, 2004).

In vitro assembly of recombinantly expressed amyloid proteins has given sustained to such findings, demonstrating that fibrils readily form when the native state is destabilized by addition of denaturant or organic solvents, low pH, high temperature, pressure, or amino- acids substitutions. In most of these conditions, the population of a partially unfolded conformation has been documented (Uversky and Fink, 2004).

However, exceptions to this empirically-derived rule have been described. For example β2-microglobulin has been found to fibrillate through a native-like folding intermediate. This amyloidogenic conformation involves the formation of a non-native trans- prolyl isomer which is part of the folding pathway of the protein, but is just rarely populated in native conditions. This suggests that the folding and aggregation landscapes of this protein are, at least, partially overlapping (Jahn et al., 2006). Similarly, an acylphosphatase variant was found to aggregate via the population of a conformation that retains its enzymatic

(22)

activity and native-like structure. Afterwards, this intermediate species interconverts slowly into enzymatically inactive amyloid-like protofibrills (Plakoutsi et al., 2005).

Figure 1.2. Landscape for protein misfolding. Analogous to the folding funnel determined for the folding of a protein, a similar landscape for the adoption of stable misfolded conformations of a protein is devised. The misfolding pathway is linked to the folding pathway through the unfolded state, and necessarily involves self- oligomerization of the polypeptide chain. In contrast to the folding situation where a single native structure is attained, several misfolded species may be populated in a stable manner, and are kinetically related. Depicted are the native and misfolded states of the amyloidogenic protein Transthyretine (native state based on 1bmz and fibril state based on 1rvs). Adapted from Jahn, 2005 (Jahn and Radford, 2005).

Limited proteolysis may also give rise to the exposure of highly aggregation prone regions, which readily self-associate, as in the process of the amyloid precursor protein APP, or in some cases with αS, giving raise to a C-terminal truncated species (Li et al., 2005). This may be explained in terms of cooperativity, another crucial factor in enabling proteins to remain soluble, ensuring that the equilibrium population of unfolded regions is minimal.

The tendency of proteins to form amyloidogenic conformations has been the subject of intense speculation by theoreticians. Electrostatic repulsion, hydrophobicity and secondary

(23)

structure propensity have all been shown to have a major influence on the tendency of fully or partially unfolded conformations to aggregate (Chiti et al., 2003). Interactions between solvent exposed, hydrophobic and flexible regions that lack stable hydrogen bonded elements of secondary structure are also suggested as factors initiating fibrillation (Uversky and Fink, 2004). Nevertheless, a specific conformational state seems not to be required for ordered aggregates to form, provided that the solution conditions permit relatively stable intermolecular interactions (Calamai et al., 2005).

1.2.3. Conformational plasticity in amyloid formation

Kinetic studies of in vitro protein aggregation have revealed a conserved fibrillation pathway for amyloid proteins, characterized by the intermediacy of ordered prefibrillar aggregates of distinct morphology (Figure 1.3) (Caughey and Lansbury, 2003). After partially unfolded states are reached, conformations rich in β-sheet secondary structure content are populated, concomitantly with oligomerization (Uversky and Fink, 2004).

Soluble oligomers rapidly convert to fibrillation intermediates, designated protofibrils, which were first described for Aβ but now seem to be a general feature of amyloid formation (Harper et al., 1997; Walsh et al., 1997; El-Agnaf et al., 2001; Haass and Steiner, 2001;

Bucciantini et al., 2002; Poirier et al., 2002). Early formed metastable protofibrils typically comprise 10 to 50 monomers, depending on the protein involved, and appear to be spherical in nature. These spheres may anneal to form chainlike protofibrils, which can further form annular pore-like species or proceed to amyloid fibrils (Harper et al., 1999; Ding et al., 2002;

Lashuel et al., 2002a; Lashuel et al., 2002b; Nichols et al., 2002).

Cellular studies with several amyloidogenic proteins suggest that the pathogenic species on this kind of disorders is an ordered oligomeric intermediate and not the fibrilar end-product of the protein aggregation pathway (Caughey and Lansbury, 2003). In particular for Aβ and αS, a possible link between a subpopulation of circularized chain-like protofibrils, the so-called amyloid pores, and neuronal death, has been proposed (Lansbury, 1999;

Goldberg and Lansbury, 2000; Haass and Steiner, 2001; Lashuel et al., 2002b). The toxicity of protofibrils is also supported by studies demonstrating that antibodies raised against soluble Aβ oligomers are capable of blocking the toxicity of these fractions to cultured neurons (Lambert et al., 2001).

It appears that the toxicity of protofibrils is related to their structure, not their sequence, as protofibrilar material comprising unrelated proteins showed similar toxicity in cell culture, while amyloid fibrils were non-toxic (Bucciantini et al., 2002; Fezoui and

(24)

Teplow, 2002). The generic nature of such aggregates and their effects on cells has recently been supported by the development of antibodies that can cross-react with protofibrills of different amyloidogenic peptides and proteins, and moreover inhibit their toxicity (Kayed et al., 2003).

Figure 1.3. Conformational plasticity in amyloid formation. A. Different conformational assemblies that may be populated during the amyloid conversion of a protein. From a completely unfolded monomer to a partially folded, β-sheet rich, monomeric or dimeric intermediates, up to oligomers and protofibrills, a extensive variety of states are populated by a protein until it reaches the formation of an amyloid fibril. B. Ideal traces for the time-dependent occurrence of the above mentioned species during aggregation. Monomer consumption (blue) is rapid, as oligomeric conformations are populated (orange). Depending on the stability of such high molecular species, oligomers coalesce into protofibrils which mature into amyloid fibrils (violet). Simultaneously, monomeric species are able to directly aggregate into fibrils (direct monomer addition), but monomer consumption is not complete as a critical concentration is always in equilibrium with the fibrils. C. Simplified nucleation- polymerization kinetic model for amyloid formation, considering a nucleation event (Knucleation) and a polymerization, or monomer addition, event (Kelongation and Kshortening).

(25)

The search for the pathogenic conformation has practical implications because its formation and the proteins with which it interacts are potential targets for therapeutic intervention. Although the protofibril hypothesis has displaced the amyloid hypothesis, controversy still exists as to which is the pathogenic species. As recently suggested by Lansbury (Caughey and Lansbury, 2003), the “culprit” should have the following properties:

- “(i) its stability and/or the rate of its formation should be sensitive to disease- associated mutations;

- (ii) it should also be accessible to the wild-type protein because most of these diseases are predominantly sporadic;

- (iii) it should be linked to a potential pathogenic mechanism (for example, sequestration of a critical protein, disruption of a membrane, or initiation of apoptosis);

- (iv) its formation/stability should be sensitive to the effects that distinguish the various cell types in the brain, since all of these diseases are selective with respect to the anatomical distribution of neurodegeneration; and

- (v) its formation/stability should be sensitive to cellular defense mechanisms that may play a role in determining disease susceptibility, such as heat shock, chaperones, proteasomal degradation, or aggregosome formation”.

At first glance it is very unlikely that a single toxic species could account for such diverse features. However, a concerted action of the different intermediates in the amyloid formation, each of them challenging a different cellular target, may reconcile the current disparate views of these sorts of afflictions.

(26)

1.3. Protein misfolding in Parkinson’s disease

Parkinson’s disease (PD) is a progressive, neurodegenerative, and age-related movement disorder affecting more than 1% of the population over 65 years of age (Goedert, 2001). It arises from the loss of dopaminergic neurons in the substantia nigra pars compacta of the brain and is accompanied by the presence of eosinophilic intracellular inclusions known as Lewy bodies (LBs) and Lewy neurites, which are confined largely to nerve cells (Galvin et al., 1999). LBs are also found in other major neurodegenerative disorders that occur increasingly with aging, including dementia with LBs and Alzheimer’s disease (Dawson and Dawson, 2003). Ultrastructurally, LBs are protein rich depositions in the form of long amyloid-like fibrils, the major component of which is the pre-synaptic protein α-synuclein (αS) (Spillantini et al., 1997). The second most common component of LBs is ubiquitin (Ub) (Forno, 1996) and other proteins that are predominantly found are heat shock proteins, neurofilaments, Tau, synphilin-1 and tubulin, among others (Shults, 2006).

The etiology of PD is not yet fully understood, but genetic analysis, neuropathologic investigations, and experimental models of PD have provided fundamental insights into its pathogenesis (Figure 1.4). Impairment of mitochondrial complex I, leading to an increase of reactive oxygen species, plays a central role in the pathogenesis of sporadic PD (Jenner and Olanow, 1998; Schapira et al., 1998; Sherer et al., 2002). PD models based on MPP+

administration feature loss of dopaminergic neurons due to inhibition of complex I with the presence of intracellular inclusions strongly immunoreactive for αS and Ub (Vila et al., 2000), and other complex I inhibitors such as paraquat or rotenone also produce symptoms of Parkinson in various animal models (Thiruchelvam et al., 2000; Manning-Bog et al., 2002;

Sherer et al., 2003).

In addition to oxidative stress, impairment in the ubiquitin-proteasomal system contributes to the disease condition in PD (McNaught et al., 2001), as has been linked to many neurodegenerative disorders (Ciechanover and Brundin, 2003). Failure of the UPS to adequately remove misfolded or abnormal proteins may underlie demise of nigral cells in sporadic PD (McNaught et al., 2001). Furthermore, deficits in the 26/20S proteasome pathways are accompanied by protein accumulation and aggregation, which may also cause neurodegeneration (Chung et al., 2001), and recently it has been demonstrated that soluble aggregated proteins can inhibit the UPS (Bence et al., 2001). These combined pieces of evidence have attracted much attention as they imply that LBs could originate from ubiquitin-

(27)

rich aggregosomes that the proteosomal components may not be able to process (McNaught et al., 2002).

Figure 1.4. Molecular pathways to Parkinson´s disease. PD is a complex disorder that involves impairment of cellular homoestasis a various levels. Three main events may trigger PD, (i) damage to mitochondria accompanied by an increase in reactive oxygen species, (ii) impairment of the ubiquitin proteasomal system, and (iii) misfolding of αS. Evidence suggests that all of these disease-associated mechanisms are tightly interconnected in a single pathological pathway that cause the demise of dopaminergic neurons in the brain stem. Adapted from Dawson, 2003 (Dawson and Dawson, 2003).

1.3.1. α-Synuclein aggregation is linked to Parkinson’s disease

Although PD is primarily a sporadic disorder, more than 10 different loci are responsible for rare Mendelian forms of PD, and the study of these gene products has provided new insights that assisted experimental models of neurodegeneration (Dawson and Dawson, 2003).

αS has been unequivocally linked to PD due to the discovery of genetic mutations.

Three different missense mutations in the αS gene result in early onset PD (A30P, E46K and A53T), and additionally, a locus triplication causing an increased dosage of the wild type (wt) αS gene potentiates the disease (Polymeropoulos et al., 1997; Kruger et al., 1998; Singleton et al., 2003; Zarranz et al., 2004). The role of αS in the formation of Lewy bodies and the pathogenesis of PD has been compared to the role of Aβ peptide and amyloid plaques in Alzheimer’s disease.

The appealing hypothesis for Lewy body formation is that αS monomers combine to form oligomers (or protofibrils), which coalesce into fibrils and then co-aggregate with other proteins into Lewy body inclusions (Conway et al., 1998; Wood et al., 1999). While the

(28)

monomers and oligomers of αS are soluble, the fibrils and Lewy bodies are insoluble in the neuronal cytoplasm. Some controversy arises from the roles of the various physical forms of αS in PD pathogenesis. LBs have been proposed to be both neurotoxic (El-Agnaf et al., 1998;

Giasson and Lee, 2001), and protective (Rochet et al., 2000; Mouradian, 2002). Other hypotheses state that the protofibrilar intermediates, made of αS oligomers, are the main species toxic to dopaminergic neurons (Conway et al., 2000). Lansbury and co-workers have recently shown that αS protofibrils can form elliptical or circular amyloid pores in cell membranes (Lashuel et al., 2002b; Volles and Lansbury, 2002), and cell culture studies found that they reduce cell viability, disrupt lysosomes and induce Golgi fragmentation (Stefanis et al., 2001; Gosavi et al., 2002). In line with these findings, the A30P and A53T mutants of αS share an increased tendency to form soluble oligomeric intermediates, whereas the E46K and A53T mutants fibrillate faster than the wt protein (Conway et al., 2000; Choi et al., 2004).

Other important genetic mutations suggest the involvement of UPS in PD, and major interest arouse with the identification of mutations in the E3 ubiquitin ligase Parkin as a cause of autosomal recessive PD (Kitada et al., 1998). Both the loss of E3 activity and the possibility of incomplete or aberrant ubiquitination are proposed as causes of Parkin-related PD (Giasson and Lee, 2003). A second member of the UPS involved in PD is the Ubiquitin Carboxy-Terminal Hydrolase-L1 (UCH-L1), and mutations in the uch-l1 gene cause dysfunction of this enzyme and lead to accumulation of toxic products (Leroy et al., 1998).

The shared symptoms of PD associated with the different genetic mutations raises the possibility of a molecular intersection in the pathogenic mechanisms driven by both protein degradation and aggregation. One plausible mechanism would involve αS mutations populating abnormal protein conformations and overwhelming the cellular protein degradation systems, whereas mutations in the UPS machinery would challenge the cell’s ability to detect and degrade misfolded proteins that can result in the formation of toxic early aggregates (McNaught et al., 2002) . The common outcome of this failure at different levels is thus expected to be a cellular buildup of such unwanted toxic species that should have been cleared in healthy conditions. Minimal defects in the crucial protein turnover machinery may suffice to cause a slow demise of dopaminergic neurons, which may explain the relentless, progressive nature of the disease (Vila and Przedborski, 2003).

(29)

1.3.2. The physiological role of α-Synuclein

αS is member of a closely related group of brain-enriched proteins (αS, βS and γS) that have been implicated in neurodegenerative disorders and cancer, among others (Clayton and George, 1998). αS and βS are predominantly cytosolic proteins and co-localize in presynaptic nerve terminals, close to synaptic vesicles, while γS appears to be axonal and cytosolic (Buchman et al., 1998).

Despite the involvement of αS in neurodegenerative diseases, the normal biological function of this family of proteins is still unclear. Expression of both αS and βS is minimal in the embryonic rat brain, thus the function of synucleins is more relevant for the adult than for the developing nervous system (Clayton and George, 1999). They are related to presynaptic terminal function, although not being vesicle proteins, yet being able to influence vesicle fusion and membrane trafficking at the presynaptic terminal (Clayton and George, 1998). In particular, αS is proposed to regulate dopamine neurotransmission by modulation of vesicular dopamine storage (Lotharius and Brundin, 2002).

An overall role of αS on brain neurotransmitter function is still elusive, since the absence of αS in mice appears associated with only mild behavioral changes. Mice lacking αS show only subtle changes in dopamine-dependent behavior and are not impaired in spatial learning (Abeliovich et al., 2000; Chen et al., 2002). In addition, mice deficient in either αS or βS, or lacking both proteins are viable, fertile, and display no major phenotype (Chandra et al., 2004). Nevertheless, αS overexpression rescues lethality associated with the lack of CSPα, a co-chaperone (HSP40 kind) associated with synaptic vesicles and implicated in folding of SNARE proteins, suggesting that αS may act as an auxiliary chaperone preserving the function and integrity of the synapse (Chandra et al., 2005).

1.3.3. The native unfolded state of α-synuclein

αS is only 140 amino acid long and lacks a defined secondary structure in aqueous solution, populating random-coil conformations (Weinreb et al., 1996). Hence it belongs to the group of “natively unfolded” or “intrinsically disordered” proteins (Uversky et al., 2000). Three distinct regions are identified in the αS polypeptide sequence (Figure 1.5):

(i) the N-terminal domain (residues 1–60), which is positively charged and contains five imperfect repeats of the highly conserved KTKEGV motif, (ii) the central NAC domain (residues 61–95), which is highly hydrophobic

(30)

(iii) The C-terminal region (residues 96–140), which is very acidic, containing 10 Glu, 5 Asp and 5 Pro residues.

Figure 1.5. Structural features associated with the primary sequence of αS. From top to bottom: Domain organization (N-terminus, NAC and C-terminus) and amino acid sequence of αS. Nature of residues colored as follows: blue (+), red (-), yellow (hydrophobic), light blue (Pro). Hydrophilicity plot (Kyte-Doolitle) calculated with a 9 residue window. Secondary structure propensity calculated using TANGO (red line α-helix, green line β-sheet, blue bars β-turns). Aggregation propensity according to TANGO (Fernandez-Escamilla et al., 2004).

Due to the despair nature of its constitutive residues, each region on αS displays a different feature associated with the disease-state. The N-terminus and the NAC region, up to residue 87, increase in α-helical structure upon association with lipid vesicles, adopting a

(31)

broken amphypatic helix, important for the physiological state of the protein (Davidson et al., 1998; Jensen et al., 1998; Jo et al., 2000; Narayanan and Scarlata, 2001; Bussell and Eliezer, 2003; Chandra et al., 2003; Ulmer and Bax, 2005; Ulmer et al., 2005). The NAC region, due to its hydrophobic nature is proposed to initiate aggregation, and indeed includes a 12 residue stretch essential for fibril formation (Giasson et al., 2001). The C-terminus of the protein is responsible for the high thermostability of the protein and is essential for the chaperone function of αS by serving as a solubilizing domain (Kim et al., 2000; Souza et al., 2000; Kim et al., 2002; Park et al., 2002b; Park et al., 2002a). Upon αS binding to lipids the C-terminus remains disordered and does not interact with the micelles (Bussell et al., 2005; Ulmer et al., 2005). Moreover, it regulates aggregation of αS since C-terminally truncated fragments are found to aggregate faster than the full-length protein (Crowther et al., 1998; Serpell et al., 2000; Murray et al., 2003). This later finding correlates with recent evidence from PD patients, in whom a C-terminal truncated αS was isolated from brain-deposited fibrilar material (Li et al., 2005).

1.3.4. Ligand induced aggregation of α-synuclein

The idea that early steps in neuronal damage by PD are governed by the oligomerization of αS is leading a number of groups to identify factors that induce a conformational change in the αS molecule. Protein and other ligand-induced conformational changes on the native conformation of the polypeptide chain might have dramatic effects on its ability to interact with itself.

Indeed, various molecules and solution conditions of manifold chemical nature have been shown to accelerate αS aggregation such as low pH, increased temperature (Uversky et al., 2001b), certain metal cations (Uversky et al., 2001c), pesticides (Uversky et al., 2001a), oxidized catecholamines (Conway et al., 2001), phospholipid vesicles, detergents and organic solvents (Perrin et al., 2000; Perrin et al., 2001). The aggregation promoting effects are suggested to be caused by a preferential population of a partially folded αS conformation which constitutes a critical intermediate on the aggregation pathway (Uversky et al., 2001b).

1.3.5. Polycation-induced α-synuclein fibrillation

The transitions of monomeric αS toward oligomeric amyloid structures have been extensively investigated in the laboratory where this thesis work was carried out. The diversity of aggregate morphology upon variation in solution conditions was studied

(32)

decrease from 7.0 to 4.0 or upon addition of polycations, fibril networks and amorphous aggregates are observed. Those aggregates retain characteristics of amyloid and presumably share a common nucleation event with the ordered fibrilar structures (Hoyer et al., 2002). The impact of the C-terminal region on the aggregation kinetics was assessed by comparing the full length with C-terminally truncated αS under solution conditions affecting the C-terminal charge state. The solution dependence of aggregate morphology was attributed to charge shielding at the C-terminal region comprising the amino acids 109-140 (Hoyer et al., 2004).

Moreover, the modulatory effect of solution conditions on A30P and A53T disease-related mutants of αS was investigated. The aggregation kinetics of wild-type αS, A30P, and A53T were modulated differently by changes of the solution conditions, suggesting long range interactions between the C-terminus with the central and N-terminal parts of the protein (Hoyer et al., 2004).

In addition to the above-mentioned conditions, work performed by our group recently demonstrated that natural polyamines promote the aggregation of αS (Antony et al., 2003).

Polyamines are cellular stabilizers of nucleic acids and membranes, being essential for growth and differentiation. In the central nervous system, they mediate channel and receptor gating, immune responses to infection (Soulet and Rivest, 2003) and are involved in neurodegenative disorders as AD (Morrison and Kish, 1995). At high intracellular levels spermidine and spermine are toxic (Auvinen et al., 1992); they are significantly elevated in the red blood cells of AD and PD patients (Gomes-Trolin et al., 2002) and can produce oxidative intermediates during polyamine retro-conversion. For these reasons, it was postulated that at physiological concentrations and in a cellular context these natural compounds may modulate the propensity of αS to form fibrils, thereby playing a significant role in the formation of cytotoxic aggregates (Antony et al., 2003).

Figure 1.6. Structure of natural polyamines accelerating αS aggregation.

The polyamines studied on this thesis work are depicted, whereas the net charge at neutral pH is consigned.

(33)

The binding site of polyamines to αS was determined by high resolution Nuclear Magnetic Resonance (NMR), and chemical shift perturbations were detected on a defined region at the C-terminal domain of the protein (Fernandez et al., 2004). This report constitutes the first evidence that residue-specific changes in this domain are linked to the acceleration of fibrillation, and is in line with previously mentioned housekeeping effect that the C-terminus has on the aggregation of αS. Binding constants were determined for the association of the different polycations to the protein, and ranged from 0.6 mM (Spermine) to 10 mM (Putrescine).

It was hypothesized that binding of polyamines might modulate structural and dynamic properties of the free protein and thereby influence intermolecular interactions required for the aggregation process. In the unliganded native conformation, the dynamic C- terminal domain could block the self-aggregating domain(s), assigned by many studies to the central core of the molecule, and C-terminal modifications such as charge neutralization by polyamine binding would release the self-aggregating domain and promote its transition to a β-sheet conformation. Further evidence supporting these findings is provided by intensity changes of the signals belonging to this domain upon polyamine binding (Fernandez et al., 2004), and by our own studies showing that C-terminally truncated αS aggregates much faster than the wt protein (Hoyer et al., 2004).

Figure 1.7. Biochemical cartoon model for the polyamine-induced aggregation of αS. The unliganded monomer is auto-inhibited for aggregation due to interactions between the NAC region and the C-terminus. Upon polyamine binding the C-terminus shielding effect is removed and the protein readily dimerizes, nucleating the aggregation process. Adapted from Hoyer et al. (Hoyer et al., 2004).

The αS-polyamine system is very suitable for study since (i) it is physiologically relevant to the pathogenesis of PD; (ii) it allows characterization of the early transitions that lead to αS fibrillation; (iii) the liganded protein is still soluble and do not readily associate at low temperature, permitting high resolution structural studies to be performed; and (iv) the charge-dependent effect on the binding affinity for the different polyamines allows a comparative assessment of the structural perturbations.

(34)

1.3.6. Metal induced α-synuclein fibrillation

Metal cations have been frequently recognized as risk factors in neurodegenerative disorders (Sayre et al., 1999; Bush, 2000). Brain lesions associated with Alzheimer’s disease are rich in Fe(III), Zn(II) and Cu(II) (Lovell et al., 1998). Recent biophysical and structural studies of the amyloid precursor protein (APP) and the amyloid-β peptide (Aβ) have provided strong evidence linking Cu(II) with AD (Atwood et al., 1998; Huang et al., 1999; Bush et al., 2003;

Brown and Kozlowski, 2004). Furthermore, a role for copper in prion disease has also been suggested and the interaction of Cu(II) with fragments of the prion protein (PrP) was structurally characterized (Viles et al., 1999; Aronoff-Spencer et al., 2000; Garnett and Viles, 2003; Brown and Kozlowski, 2004).

Although less well defined, the metallobiology of PD is attracting increasing attention. Iron deposits have been identified in Lewy bodies in the substantia nigra (Castellani et al., 2000) and elevated Cu(II) concentrations have been reported in the cerebrospinal fluid of PD patients (Pall et al., 1987). Based on this evidence, a role was suggested for copper and iron in the catalysis of oxidative oligomerization and subsequent aggregation of αS in the presence of hydrogen peroxide (Hashimoto et al., 1999; Paik et al., 2000).

A systematic analysis of the effect of various metal ions revealed that Al(III), Fe(III) and Cu(II) accelerate AS fibrillation in vitro (Uversky et al., 2001c) and another study showed that Cu(II) is the most effective ion in promoting αS oligomerization (Paik et al., 1999). The enhancement of fibrillation by metals has been attributed to Coulombic screening of charge-charge repulsion associated with the highly acidic C-terminal domain (Paik et al., 1999; Uversky et al., 2001c). Cu(II), Fe(II) and Mn(II) quench tyrosine fluorescence and are presumed to form stable metal-protein complexes with αS and cause a significant acceleration of αS fibril formation (Uversky et al., 2001c; Golts et al., 2002).

(35)

1.4. The unfolded state of proteins

As stated before, the unfolded state of a protein represents the starting conformation for the folding pathway, and partially or completely unfolded conformations are also key intermediates in amyloid formation. In addition, the proteins investigated in this thesis work are classified as natively unfolded proteins. It is thus instructive to comment at this point on the particular characteristics of this conformational state of the polypeptide chain.

1.4.1. Intrinsic conformational restrictions in the unfolded state

As originally defined by Tanford (Tanford, 1968), the conformations populated by an unfolded polypeptide chain are true random coils, retaining no element of their original native structure. Nevertheless, even in the most expanded unfolded protein sequence, local elements of the chain are held in proximity via the peptide bond, leading to the Flory isolated pair hypothesis (Flory, 1969), which states that in a random coil conformational restrictions do exist, but are limited to nearest neighbor interactions. However, it was recently suggested that systematic local steric effects can extend beyond nearest-chainneighbors and may restrict the size of accessible conformationalspace (Pappu et al., 2000). Shortle has recently restricted the definition of a random coil state of a polypeptide as the well defined reference (and perhaps ideal) state in which no sidechain-sidechain interactions occur (Shortle, 1996). Thus the characteristics of a random coil state of a protein will depend on the intrinsic molecular features of the amino acids that compose the particular polypeptide chain.

In a polypeptide chain the peptide bond is planar, and the main-chain trace of the backbone can be described in terms of peptide planes connecting the Cα atoms. The relative orientation of two of these planes (i and i + 1) will be determined by the pair of torsion angles φi and ψi, defining the rotation about the bonds connecting the Cαi (Figure 1.8).

Figure 1.8. A polypeptide chain. Schematic diagram of a polypeptide chain showing the torsion angles f and y, which determine the backbone conformation.

(36)

The combination of φ and ψ angles for each of the amino acids composing the chain defines the specific conformation of the polypeptide backbone. Thus for proteins that retain a single structure, each residue populates a defined φ, ψ conformation, with small fluctuations, contrasting a random coil conformation, where there will be a distribution of φ, ψ angles for each residue, giving rise to an ensemble of conformers (Smith et al., 1996).

Ramachandran in the late 60´s observed that steric considerations in the backbone atoms are sufficient to restrict the permissible φ, ψ values for a given amino acid in a polypeptide chain, providing the basis for the description of conformational spaces for secondary structure formation in proteins (Ramachandran et al., 1963). The Ramachandran plot depicts the distribution of φ, ψ torsion angles that are evidenced in a polypeptide and relates those values with the adoption of helical or sheet backbone conformations (Figure 1.9). Thus the measurement of torsion angles for a given residue may, in general, report on the conformation in which a certain amino acid is embedded.

Figure 1.9. Ramachandran plot and dihedral angles for secondary structure conformations in polypeptides. Diagram of the φ, ψ torsion angles distribution for residues in a polypeptide chain. The average conformational parameters are consigned for each secondary structural element.

In the absence of secondary structure, the values of φ, ψ torsion angles for a given residue will not be correlated with the φ, ψ angles for all the other residues, and will exclusively report on the conformational space that the residue is sampling (Smith et al., 1996). A database analysis of residues in coil regions of known structures of proteins, i.e.

residues not in α-helices and not in β-sheets, has determined the intrinsic residual preferences for amino acids, in the absence of interactions stabilizing secondary structural motifs. The striking conclusion from such studies is that the distribution of φ, ψ angles in coil regions of

(37)

proteins is far from random, with backbone angles heavily biased to those corresponding to α-helices and β-sheets. (Serrano, 1995; Swindells et al., 1995). Within this framework, specific residues have preferences for particular areas of conformational space, and these preferences are heavily influenced by neighboring residues (Griffiths-Jones et al., 1998).

Nevertheless, the high population of α and β space predicted for residues in a random coil does not itself imply a persistent population of α-helices or β-sheets, since such population requires repeated values (at least five) of the appropriate φ, ψ dihedral angles along a given sequence (Smith et al., 1996). Questions remain, however, regarding the length scale over which the distribution of φ, ψ torsion angles becomes uncorrelated, giving rise to the concept of persistence length of the chain, which could define some sequence-local structure (McCarney et al., 2005).

1.4.2. Structural studies on the unfolded state of proteins

Investigation of non-native states of proteins is now possible through a wide variety of techniques, which provide information on global parameters, such as the radius of gyration or sedimentation coefficient, and on local properties of the chain, like the NMR-derived constraints (Table 1.2).

The concerted use of several of these techniques is necessary, due to the very flexible nature of the chain and to the absence of a defined conformation. Different properties of the polypeptide are probed by each of these methods which provide complementary information to characterize such dynamic state. However, in particular for the study of the denatured state of proteins, it has been difficult to reconcile the results arising from many of these methods.

For example, the overall topology of the unfolded state fits the expected for a random coil polymer, deprived of conformational preferences, challenging fluorescence and NMR- derived parameters, which unequivocally show the presence of long-range and residual local structure (Smith et al., 1996; McCarney et al., 2005).

This scenario rise the important question as to what extent are the structures of unfolded proteins disordered. It also points towards how general are structural propensities in unfolded proteins, questioning whether they are sequence-specific or denaturant-induced (Baldwin, 2002). Recently, computer simulations on ensemble-averaged structural coil models have reconciled this view (Bernado et al., 2005b; Jha et al., 2005a). The presence of defined local backbone structure in unfolded proteins suggests that conformations populated under strong denaturing conditions are not as heterogeneous as previously supposed, with a

(38)

consequent reduction in the entropy of the unfolded state relative to predictions from the random coil model (Fitzkee and Rose, 2004).

Table 1.2. Methods to study conformational properties of the unfolded state of proteins. Adapted from (Smith et al., 1996; Dyson and Wright, 2004).

A considerable body of data has been generated to date on the denatured state of proteins, providing structural details on the conformations that are particularly populated. The

Referenzen

ÄHNLICHE DOKUMENTE

Despite the impairment in amyloid fibril formation in vitro, it increases neurotoxicity in primary neurons, worms, and flies (Karpinar et al. 2011), reinforcing the idea

PrP C shows a globular C-terminal domain, in which NMR spectroscopy revealed three well-defined alpha-helices and two small beta-sheets 109 (blue and yellow stretches,

in which no target protein could be detected in the soluble fraction after refolding (Fig. All remaining conditions showed soluble target protein. As with the minimal RRM

Here, I present the crystal structure of the functional important human SCOC coiled coil domain (residues 78–159), which was determined at 2.8 Å resolution.. SCOC forms a parallel

The results of these measurements were comparable to the results of the FCCS measurements on complexes assembled on Actwt pre- mRNAs (see g. 26) and indicate that even when

On the basis of the data mentioned above I propose as my conclusion that the epistle written to the Romans, 1 Corinthians, 2 Corinthians, the epistle to the Philippians and

Based on these considerations, we offer the following suggestions for selecting suitable moment closures for approximating dynamics in discrete space: (1) mean field-

If many of such excitable units are spatially coupled, they form an excitable medium, which supports the propagation of non-linear excitation waves. Due to the above local properties